[Next K-Startup] The 'AI Personal Stylist' I Create

Lee Seonjun OD CEO "Find the Right Clothes via Chat"
Developed Based on Experience from Two Previous Startups

Editor's NoteThe competition rate for the first accelerating program of Next Elevation, a startup accelerator established by Asia Economy, was 26 to 1. Over 130 startups from various fields applied, and 5 startups were finally selected. The competition is still ongoing. Startups selected as outstanding companies during the 3-month program can receive up to 100 million KRW in investment. Next Elevation plans to continue attracting follow-up investments through a demo day scheduled for December. The curtain on this heated competition has already risen. Given the fierce competition, the ambitions of the 5 startups are extraordinary. We visited the site where the forging for innovation is in full swing.

Lee Seon-jun, OD CEO, interviewed by Asia Economy regarding the 'Next Elevation' program. Photo by Jo Yong-jun jun21@

"The idea is to train artificial intelligence (AI) with my own fashion style to create a ‘personal stylist.’"

Lee Seon-jun, CEO of OD, has been interested in fashion and loved clothes since childhood. His hobby was to check out new brands and clothing designs. The company name ‘OD’ is simply the English word for ‘clothes.’ During shopping, he became curious. Looking around, he noticed many people spending time scrolling through various fashion brands on social networking services (SNS) like Instagram. He thought it would be better to find his style by chatting with AI like a search engine rather than doing that.

Lee, who studied business administration, computer engineering, and mechanical engineering at the University of California, Berkeley, decided to create such a service. He explained, "There are already chat applications combining text-based machine learning models, and while search engines cover a broad range of fields, OD is creating an app specialized in fashion. AI learns only fashion so it can understand my style one-on-one."

The biggest difference from existing online shopping malls is ‘communicating through text.’ Lee said, "When you enter existing shopping malls, the layouts are all the same. You enter as a male, click on pants, filter by desired color and fit, and find the clothes you want. What OD is developing now is a method where you type in the chat window, ‘I need black skinny jeans for Halloween,’ and it finds clothes that match." To do this, they collected size data from each brand. They plan to continuously update it to prepare for size and fit changes by season even within the same brand.

Changing tastes can also be adjusted and applied through continuous learning. When users encounter new brands or clothes, they directly rate them as ‘good’ or ‘not good’ to train the AI. This is similar to swiping right on a photo for a person you like or left to skip on a dating app.

Lee Seon-jun, OD CEO, interviewed by Asia Economy regarding the 'Next Elevation' program. Photo by Jo Yong-jun jun21@

It has been less than two months since the idea was conceived. Although the results are not out yet, Lee is focusing on development based on his experience of having created startups twice before, such as ‘Endorsify’ and ‘Minti.’ Brian Wang, a longtime friend since college and AI platform lead at SambaNova Systems, has also joined to speed up the work. SambaNova is a US-based AI semiconductor company. Lee said, "He was a friend I promised to start a business with someday, and after judging that he had gained a lot of experience, we united under OD."

The goal is to create a minimum viable product (MVP) to test user experience (UX) before the Next Elevation program ends. Lee said, "The accuracy of recommendations and chat models can be revised later, so excluding those, I want to test the MVP directly in the market for a few weeks. Even after the program, we plan to continue improving the model and developing the product."